rendered.ai Raises $6M to Solve Data Scarcity

The Challenge of Data Availability for Tech Companies

Companies striving to launch software-driven products and services often find themselves hampered by a critical obstacle: the lack of sufficient data. This issue is being addressed by Rendered.ai, a data startup founded two years ago, which specializes in the creation of synthetic data for sectors including satellite technology, healthcare, robotics, and automotive engineering.

What is Synthetic Data?

Essentially, synthetic data is created through artificial means, rather than being collected from real-world observations. As Rendered.ai CEO Nathan Kundtz clarified in a recent TechCrunch interview, their focus lies on “engineered simulated data sets,” specifically those built upon physics-based simulation.

The Origins of Rendered.ai

Nathan Kundtz holds a PhD in physics from Duke University and gained significant experience in the space sector. He previously led antenna development at Kymeta Corporation. Following his departure, he consulted with numerous emerging space companies.

It was during this period that he identified a recurring “chicken and egg” dilemma facing these businesses.

A Common Scenario: Sensor Development and Funding

Consider a company innovating a novel sensor for satellite applications seeking investment for commercialization. Demonstrating the sensor’s potential to investors requires showcasing its ability to yield valuable insights.

However, generating these insights necessitates launching a satellite constellation and accumulating substantial data – a process that is both costly and time-consuming.

The Impact on Artificial Intelligence

“This limited access to data was actively impeding the progress of artificial intelligence,” Kundtz stated.

Rendered.ai aims to resolve this bottleneck by providing readily available, high-quality synthetic datasets, enabling companies to accelerate development and secure funding without the initial hurdles of real-world data acquisition.

Investor Attention for Rendered.ai

Rendered.ai’s innovative method for expanding data accessibility has successfully attracted investor interest. The company recently secured $6 million in a seed funding round. Space Capital spearheaded the investment, with further participation from Tectonic Ventures, Congruent Ventures, Union Labs, and Uncorrelated Ventures.

A key differentiator for Rendered.ai lies in its utilization of a physics-based methodology. This sets it apart from competitors primarily employing generative techniques for synthetic data creation.

Many competing solutions rely on augmenting existing datasets through generative adversarial networks (GANs). These networks utilize competing neural networks to simulate and improve synthetic data quality. However, according to Kundtz, this approach offers limited benefit to nascent industries lacking substantial initial data.

Challenges in Data Acquisition

Acquiring sufficient data can present significant hurdles for companies. The process is often expensive, complex, and demands considerable time investment.

These difficulties are amplified when dealing with data formats beyond standard RGB imagery, such as that produced by synthetic aperture radar systems.

Leveraging Physics for Data Generation

How does a physics-based approach overcome the limitations of generating novel information? “By incorporating our understanding of physics, and the underlying equations governing phenomena like light interaction, we can introduce new information into the algorithmic creation process,” explained Kundtz.

“This allows us to simulate appearances under varying conditions and subsequently generate comprehensive datasets.”

Synthetic data generated in this manner provides a valuable resource for training and validating AI models, particularly in scenarios where real-world data is scarce or unavailable.

• The physics-based approach allows for the creation of data even when no initial data exists.
• It addresses the cost and time constraints associated with traditional data acquisition.
• It extends to non-RGB image types, like those from synthetic aperture radar.

Rendered.ai’s technology has the potential to accelerate development in industries reliant on computer vision, offering a pathway to overcome data limitations.

A Developer-Focused Toolkit

Rendered.ai has created a platform offering both a no-code configuration interface and APIs. These tools empower users to engineer and refine the settings of a given data set.

Furthermore, the platform includes functionalities for in-depth data set analysis and introspection. Rendered.ai also supplies initial code examples tailored to applications like processing satellite imagery.

This comprehensive offering is described by the company as a “platform as a service” model.

Decreasing Skill Requirements

Although utilizing Rendered.ai currently necessitates a degree of technical proficiency, Kundtz notes that this requirement is consistently diminishing. Ongoing funding will be allocated to further reduce the skill level needed for platform operation.

The goal is to enable anyone with basic browser navigation skills to generate synthetic data. This includes not only creation, but also precise control over the characteristics of the generated data and seamless integration into existing machine learning processes.

Iterative Data Refinement

Often, organizations are unaware of the specific parameters required to create effective synthetic data or optimize algorithm performance. Rendered.ai employs an iterative methodology, highlighting the platform’s interactive nature.

This interactivity allows customers to pinpoint deficiencies in their algorithms and gain a clearer understanding of potential limitations.

The Future of Data

Kundtz believes that synthetic data won’t entirely supersede real-world data. However, its role in artificial intelligence applications is expected to grow significantly.

The platform also presents an opportunity to lessen the data advantage held by large corporations, such as Google, who possess exclusive access to vast data repositories.

Platform Expansion and Investment

Rendered.ai has onboarded an initial group of customers, but the platform remains in a beta phase. The recent funding will be used to broaden platform access and invest in specialized data sets for specific industry sectors.